FAQ (import)

FAQ concerning TM import:

Is it possible to import XLIFF with its statuses without checking the option "import statuses" in the UI?

It is currently not possible, the option must be checked. However, it is possible in TMX files.

Can you change the Modify the existing TM record setting via the API? Example: there are scheduled automated imports but each import should be with different setting, can you use the API to change the setting, run the import, and then restore the regular setting?

It is only a global setting, there is no such option to perform that kind of action via API. It is possible through the database.

Is it possible to import a big file (190 MB already zipped) through UI?

In the UI there is a limit of 150 MB per file. Bigger files can be imported through API or if the client has no license for it - we can use our internal tool to split the TMX into smaller files.

If the TM has segments IDs, is each segment created as a separate TM entry, even if the source is identical?

Yes. For more information about duplicates management refer to Duplicates behavior (TM import).

If the TMX file contains target language pt_BR, but in the UI you import it as pt_PT will it import anyway?

Yes, only the first part of the language code, in this case “pt“, is taken into consideration. If you import it through the UI choosing target “Portuguese” (pt_PT) the segments will be imported as Portuguese.

Is it possible to import TM as xlsx together with all the metadata that is inside the file?

The only metadata possible to be imported via Excel is segment id, the column containing it must be named “segmentid“. Other metadata will not be imported, but xlsx files can be converted to TMX through our internal tool. In this case please write to help-localization-engineers-migration channel on Slack.

Is the imported TM stored anywhere in backend?

It is available for a limited period in this path /xtm/xtm/temp/tmx, its name starts with “upload” and has the same name that it had in the import logs.

Where do language codes have to be accurate? File name, within the file, other?

For both API import and UI import the language codes need to be accurate (list of all codes is available in XTM Manual) in the request itself. Inside the file they must be placed as an attribute of tuv and it is enough if they have the first part of the language code. Example: you want to import TM source en_US, target zh_CN and you have already chosen the language in the UI. It is enough if inside the file the source is set as en and target as zh. The most important thing is to set it correctly in the request. The file will not import at all if inside of it there are completely different language codes than what you choose in the request.

File name does not matter at all.